1. Field of the Invention
The present invention relates to a control device having an error correction function which is used in a memory and communication. By the error correction function, an error is automatically corrected so as to enhance reliability of a device and a system. Further, the present invention especially relates to a control device which has an error correction function and is used in industrial device and system which have to be normally operated under an inferior operation environment due to temperature, vibration, noise, or the like and a device and a system whose reliability is required such as medical, aerospace, and public systems.
2. Description of the Related Art
Failure prediction performed by using a device lifetime which represents a period in which a device can be normally used enables advance preparation of replacement of components so as to be able to shorten failure recovery time. Shortening of failure recovery time improves capacity of a production line, so that it is required to predict device lifetimes of devices on a production line with high accuracy.
In general, failures of specific parts are increased along with secular change in a memory in which writing and deletion of information are repeated. A system constituted of a memory having an error correction function operates without any problems in a case of an occurrence of a failure within a range of correction capability of the error correction function. On the other hand, in a case where a failure beyond the correction capability occurs, correction is impossible and the system has to be stopped.
Therefore, an operation rate of the error correction capability is set to a device lifetime so as to be able to prepare a countermeasure against a failure before the system stops due to a failure of a memory.
As a prior art technique, a failure prediction method in which deterioration progress of a device is calculated based on a correction record of an error correction function so as to take measures against a failure in an early stage is proposed (refer to Japanese Patent Application Laid-Open No. 2012-155737).
Here, the error correction function is an error correction method by which an error is automatically corrected by using an error correction code so as to improve reliability of data. The error correction code has been conventionally studied as a code logic and various structures of correction codes having different functions such as a correction code specialized for a burst error and a correction code whose capability of error correction is locally enhanced have been proposed.
As a prior art technique, an error correction function in which a plurality of error correction codes are provided and the error correction codes are switched is proposed (refer to Japanese Patent Application Laid-Open No. 2012-178014).
In general, in design of an error correction function, an error correction code having enough tolerance is designed while assuming an error generated by an error cause depending on variation in manufacturing of a device, an installation environment, and an operation state. Therefore, an error correction function using a single error correction code is a waste of resources in a certain individual.
There are a plurality of forms in a failure of a device. Further, a cause accelerating a failure of a device varies depending on the form. For example, in a memory device, there is a case where probability of an occurrence of a failure in which a specific memory region is fixed on a specific state due to unevenness of the quality or the like is increased when the memory device is used in a high temperature environment. In general, a cause of a failure is grown resulting from an operating state of a device. Therefore, if a causal relationship between a failure of a device and an operating state of the device is known and can be reflected, a device lifetime can be predicted with high accuracy.
It is possible to provide a control device which has a function to calculate a device lifetime highly accurately by reflecting a result of failure analysis of every operating state to calculation of the device lifetime and outputs a maintenance preparation request and a maintenance execution request in accordance with a current device lifetime in an error correction function in which error correction codes are switched in accordance with an operation state.
In order to do that, a device lifetime may be able to be first calculated by using an error correction function in which error correction codes are switched. However, a device lifetime which is calculated based on the number of correctable bits in a prior art technique is discontinuous due to switching of an error correction code in an error correction function for switching to an error correction code having the different number of correctable bits. Thus, it has been impossible to correspond to switching of error correction codes.